Saturday, June 20, 2009

Rotisserie Baseball and the Replacement Player

Earlier this season as I was kvetching about non-Patton Player Values (as I'm wont to do), both Rotoman and Brett talked a little bit about how "replacement level" players are factored in to some Rotisserie pricing systems.
Like Rotoman said, maybe they calculate replacement level on a position-by-position basis (and they should)...
Instinctively, I have a problem with the idea of the "replacement" level player in Rotisserie and how it impacts (or, more accurately, doesn't impact valuation). But as a starting point for this article, I thought I'd take a look at how valuable a representative pool of free agents were position by position.

Billy Almon Brown Graduate 2007-2008
Free Agent Value by Position

Pos
Year
Avg Value
CO
2007
$6.0
OF
2007
$5.3
OF
2008
$5.3
MI
2008
$4.8
CO
2008
$3.1
MI
2007
$2.0
C
2007
$0.9
C
2008
$0.5

While outfielders are valuable, they weren't the most valuable in my American League the last two years. That honor would go to the corners in 2007. But that's because Carlos Pena was available via FAAB that year. Take Pena out and the average CO value plummets to $2.4 per player.

So Pena's an anomaly, one could argue. But isn't he, by definition, a "replacement" player? Yes (at least in this league). So he should count.

In any event, how do I calculate "replacement-level" value for Rotisserie? If I had used 2007's average middle infield replacement values to do so, I would have overpaid for middle infielders in 2008...especially considering that Mike Aviles (earned $19), Jamey Carroll ($8), Willie Bloomquist ($7), and Brendan Harris ($7) were coming along.

I guess I could decide to discount outfielders because it's more likely that I'm going to get a $5 replacement level player from year to year than I am a corner or middle infielder. But while I know this with certainly at the end of the season, I don't during the season. Franklin Gutierrez ($11), Victor Diaz ($5), and Josh Rabe (-$1) all went for the same FAAB bid on the same week in 2007. I'm a little uncomfortable discounting stats too much for my outfield pool "knowing" that Gutierrez is coming along since I don't know that it's Gutierrez coming along but an OF who will earn $11.

Is there some merit to the idea of paying certain players more based on position? Probably. Is it something that can be measured? Perhaps, but I believe there is too much fluctuation from year to year in deep leagues to offer a reliable yardstick.

1 comment:

Brett said...

"Is there some merit to the idea of paying certain players more based on position? Probably. "

Of course there is position scarcity - you know this. We just don't worry about it much because it only matters for catchers, and even then, usually only to the tune of about $2 (the replacement level catcher is usually about $2 worse than replacements at other positions, so you have to add $2 to every catcher's value).

In theory, it exists at other positions too, but the values are so small that it's not worth worrying about.

In a shallow mixed league, though, you really do need to calculate it at every position. If I recall correctly, catchers are worth about $10 more than first basemen, for example. And 2nd basemen maybe $5 more.


And about what replacement level is - it's the value of a replacement that you could get for free. That's obvious, but the point is, when you do your preseason projections, you assume they are 100% correct. You then look at your top 24 catchers (how you decide they are the top 24 without having their values yet is a bit of a paradox, but you can usually just do it intuitively, and if it's that hard to decide who is #24 and who is #25, it probably doesn't matter anyway). Then replacement level is (strictly speaking) the stats of #25. You might want to use an average of #23-#27 or something to smooth things out (more important with, say, middle infielders, where #25 could have all his value wrapped up in SB, where #26 could have 4 home runs).


I think you're looking at the replacement level players the wrong way in your chart above - Carlos Pena who hit 46 HR in 2007 is not replacement level. Carlos Pena who was projected to hit .220 with 8 HR is - because you were doing this preseason. And if you're doing it at the end of the season, Pena is clearly one of the best players and some other scrubs make up replacement level.

That's the way I see it at least - things are pretty much static before the auction, where you're relying solely on your projections. Once the season starts, everything's in flux.